Joint hyperbolic and Euclidean geometry contrastive graph neural networks

Graph Neural Networks (GNNs) have demonstrated state-of-the-art performance in a wide variety of analytical tasks. Current GNN approaches focus on learning representations in a Euclidean space, which are effective in capturing non-tree-like structural relations, but they fail to model complex relati...

Full description

Saved in:
Bibliographic Details
Main Authors: XU, Xiaoyu, PANG, Guansong, WU, Di, SHANG, Mingsheng
Format: text
Language:English
Published: Institutional Knowledge at Singapore Management University 2022
Subjects:
Online Access:https://ink.library.smu.edu.sg/sis_research/7564
https://ink.library.smu.edu.sg/context/sis_research/article/8567/viewcontent/Joint_Hyperbolic_and_Euclidean_Geometry_Contrastive_Graph_Neural_Networks_revision_version.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Singapore Management University
Language: English
id sg-smu-ink.sis_research-8567
record_format dspace
spelling sg-smu-ink.sis_research-85672024-01-09T01:19:49Z Joint hyperbolic and Euclidean geometry contrastive graph neural networks XU, Xiaoyu PANG, Guansong WU, Di SHANG, Mingsheng Graph Neural Networks (GNNs) have demonstrated state-of-the-art performance in a wide variety of analytical tasks. Current GNN approaches focus on learning representations in a Euclidean space, which are effective in capturing non-tree-like structural relations, but they fail to model complex relations in many real-world graphs, such as tree-like hierarchical graph structure. This paper instead proposes to learn representations in both Euclidean and hyperbolic spaces to model these two types of graph geometries. To this end, we introduce a novel approach - Joint hyperbolic and Euclidean geometry contrastive graph neural networks (JointGMC). JointGMC is enforced to learn multiple layer-wise optimal combinations of Euclidean and hyperbolic geometries to effectively encode diverse complex graph structures. Further, the performance of most GNNs relies heavily on the availability of large-scale manually labeled data. To mitigate this issue, JointGMC exploits proximitybased self-supervised information in different geometric spaces (i.e., Euclidean, hyperbolic, and Euclidean-hyperbolic interaction spaces) to regularize the (semi-) supervised graph learning. Extensive experimental results on eight real-world graph datasets show that JointGMC outperforms eight state-of-the-art GNN models in diverse graph mining tasks, including node classification, link prediction, and node clustering tasks, demonstrating JointGMC's superior graph representation ability. Code is available at https://github.com/chachatang/jointgmc. (c) 2022 Elsevier Inc. All rights reserved. 2022-09-01T07:00:00Z text application/pdf https://ink.library.smu.edu.sg/sis_research/7564 info:doi/10.1016/j.ins.2022.07.060 https://ink.library.smu.edu.sg/context/sis_research/article/8567/viewcontent/Joint_Hyperbolic_and_Euclidean_Geometry_Contrastive_Graph_Neural_Networks_revision_version.pdf Research Collection School Of Computing and Information Systems eng Institutional Knowledge at Singapore Management University Graph neural networks Hyperbolic embedding Contrastive learning Graph representation learning Databases and Information Systems Graphics and Human Computer Interfaces OS and Networks
institution Singapore Management University
building SMU Libraries
continent Asia
country Singapore
Singapore
content_provider SMU Libraries
collection InK@SMU
language English
topic Graph neural networks
Hyperbolic embedding
Contrastive learning
Graph representation learning
Databases and Information Systems
Graphics and Human Computer Interfaces
OS and Networks
spellingShingle Graph neural networks
Hyperbolic embedding
Contrastive learning
Graph representation learning
Databases and Information Systems
Graphics and Human Computer Interfaces
OS and Networks
XU, Xiaoyu
PANG, Guansong
WU, Di
SHANG, Mingsheng
Joint hyperbolic and Euclidean geometry contrastive graph neural networks
description Graph Neural Networks (GNNs) have demonstrated state-of-the-art performance in a wide variety of analytical tasks. Current GNN approaches focus on learning representations in a Euclidean space, which are effective in capturing non-tree-like structural relations, but they fail to model complex relations in many real-world graphs, such as tree-like hierarchical graph structure. This paper instead proposes to learn representations in both Euclidean and hyperbolic spaces to model these two types of graph geometries. To this end, we introduce a novel approach - Joint hyperbolic and Euclidean geometry contrastive graph neural networks (JointGMC). JointGMC is enforced to learn multiple layer-wise optimal combinations of Euclidean and hyperbolic geometries to effectively encode diverse complex graph structures. Further, the performance of most GNNs relies heavily on the availability of large-scale manually labeled data. To mitigate this issue, JointGMC exploits proximitybased self-supervised information in different geometric spaces (i.e., Euclidean, hyperbolic, and Euclidean-hyperbolic interaction spaces) to regularize the (semi-) supervised graph learning. Extensive experimental results on eight real-world graph datasets show that JointGMC outperforms eight state-of-the-art GNN models in diverse graph mining tasks, including node classification, link prediction, and node clustering tasks, demonstrating JointGMC's superior graph representation ability. Code is available at https://github.com/chachatang/jointgmc. (c) 2022 Elsevier Inc. All rights reserved.
format text
author XU, Xiaoyu
PANG, Guansong
WU, Di
SHANG, Mingsheng
author_facet XU, Xiaoyu
PANG, Guansong
WU, Di
SHANG, Mingsheng
author_sort XU, Xiaoyu
title Joint hyperbolic and Euclidean geometry contrastive graph neural networks
title_short Joint hyperbolic and Euclidean geometry contrastive graph neural networks
title_full Joint hyperbolic and Euclidean geometry contrastive graph neural networks
title_fullStr Joint hyperbolic and Euclidean geometry contrastive graph neural networks
title_full_unstemmed Joint hyperbolic and Euclidean geometry contrastive graph neural networks
title_sort joint hyperbolic and euclidean geometry contrastive graph neural networks
publisher Institutional Knowledge at Singapore Management University
publishDate 2022
url https://ink.library.smu.edu.sg/sis_research/7564
https://ink.library.smu.edu.sg/context/sis_research/article/8567/viewcontent/Joint_Hyperbolic_and_Euclidean_Geometry_Contrastive_Graph_Neural_Networks_revision_version.pdf
_version_ 1787590747818557440